A Wiener filter-based deconvolution algorithm is developed to restore vibration-degraded video imagery from an intensified CCD camera. The method is based on the use of azimuth and elevation angular optical line-of-sight data recorded from external sensors to estimate a two-dimensional vibration-blur impulse response on a per frame basis. Flight conditions are reproduced in the laboratory by use of prerecorded in-flight vibration data. The performance of the algorithm varies from frame to frame, following the time-varying characteristics of the vibration-blur impulse response. However, real-time display of the restored video minimizes these effects because of eye integration, and near-full restoration of the original uncorrupted imagery is observed for both high-light- and low-light-level conditions with minimal amplification of noise.
展开▼
